Eigen.Spaces:
Accurate and timely risk measurement has always been fundamental to capital markets activity. With markets moving into uncharted territory with unprecedented speed, the need for timely risk measurement has never been so critically important.
Approximations that work well in normal times to provide real time risk estimates are no longer useful; today’s extreme conditions routinely violate many assumptions that underpin their validity. Adding more hardware yields some performance gains, but when risk analytics are not optimized for parallel computation, a lot of potential is left unexploited.
Re-tooling analytics to utilize the parallel performance available in multi-core processors has been an expensive proposition. Until now.
eigen.spaces, a performance toolkit based on Intel Thread Building Blocks is set to change all that. Allowing rapid integration of tested and familiar analytics into a high performance parallel computation framework, it realizes the full potential of existing valuation models with multi-core processors and high performance server grids.
Used within electronic trading platforms, the performance gain translates to larger flows, better execution and lower latency - a crucial competitive edge. Algorithmic strategies can increase in sophistication without losing valuable milliseconds, arbitrage threats can be eliminated and new opportunities exploited.
Eigen.Spaces: eigen.spaces is a C++ concurrency toolkit utilizing Intel’s Thread Building Blocks and low latency network protocols, aimed at easy and economic deployment of serial, thread-safe analytics. In addition to concurrency components, eigen.spaces provides building blocks for adaptive load management and fail-safe operation.
Exploit the inherent concurrency in financial computation while preserving significant software investment, and minimize hardware costs by more efficient utilization of processing capacity.
Bring existing computationally intensive tasks into real-time, high frequency environments. Gain consistency between spreadsheet and mainframe risk computation by sharing common analytics code and market data.
The load management algorithms embedded in eigen.spaces are designed to minimize operations monitoring and management costs, while providing real time performance statistics to support continuing refinement and performance improvement.
download eigen.spaces pdfeigen.spaces | red.shift | model.bricks